Академический Документы
Профессиональный Документы
Культура Документы
i
/ qR
T
,
i
is the absorption cross section of the B state
i =1 or M-state i =2,
i
is the quantum yield of photoi-
somerization from BM state i =1 and MB i =2 and
is the angular frequency of the light. From the Eq. (1) it
follows that the dependence of both the grating contrast and
decay time of diffraction signal on reference beam intensity I
is in agreement with our experimental observation. In addi-
tion, our experimental results of the diffraction decay can be
explained in terms of two lifetimes. For example, theoretical
plot (solid line) in Fig. 2 is obtained using two lifetimes, 62
and 240 ms. We believe the short one originates from the
isomerization rate from MB, and the long one is due to the
mechanism of bR molecular reorientation. We modied Eq.
(1) with two exponential decays corresponding to these two
lifetimes to t the experimental data in Fig. 2. Diffraction
decay data for other object intensities can also be explained
similarly.
The intensity in Fourier spectrum of any object is spa-
tially distributedlow frequencies are at high intensity and
high frequencies are at low intensity. We exploit the temporal
and intensity dependent features of diffraction efciency of
bR sample for selective recording of spatial frequencies for
applications in image processing. To illustrate the feasibility
of the technique we rst recorded high spatial frequency
components of a binary object E, by matching the corre-
sponding intensity to the intensity of the reference beam. The
recording process takes about 5 s so that the contrast of Fou-
rier hologram reaches its maximum in the bR lm. When the
object beam is blocked, the reconstructed wave by the refer-
ence beam shows an edge enhancement effect. The experi-
mental results are shown in Fig. 3(b). In addition, the bR
material also has saturable absorption feature which enables
it to directly pass low-frequency components of the image
and block high-frequency components. In the absence of ref-
erence beam, an image with soft edge of object E as shown
in Fig. 3(c) is obtained. Thus this scheme is able to give the
processed results of both low-frequency and high-frequency
components in real time and can be adopted for other fre-
quency bands as well.
We now exploit this scheme for processing mammo-
grams for detection of microcalcications. The microcalci-
cations (tiny calcium deposits in human breast) correspond
to high spatial frequencies in the Fourier spectrum because
of their small size 10 m and diffuse nature. The object
E in the above experiment is replaced with region of interest
(ROI marked by the radiologist) of the mammogram as de-
picted in Fig. 3(d). By controlling the reference intensity,
edge enhancement of mammograms is achieved leading to
FIG. 2. Experimental results of temporal decay of diffraction signal at
matched object and reference beam intensity of 26 mW/ cm
2
. The solid line
is the theoretical t. Inlet: diffraction signal as a function of object beam
intensity for various time delays with xed reference beam intensity of
26 mW/ cm
2
.
Appl. Phys. Lett., Vol. 85, No. 24, 13 December 2004 Kothapalli et al. 5837
Downloaded 29 Mar 2012 to 158.121.194.143. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
clear display of microcalcications (not visible to the naked
eye of the radiologist) on the monitor as shown in Fig. 3(f).
Irrespective of different size of regions of interest and den-
sity of soft tissue present in mammograms, the radiologist
can easily scan for desired microcalcication clusters by ro-
tating the variable attenuator placed in the path of the refer-
ence beam. With the spatial resolution of bR lm
6000 lp/ mm
11
and CCD camera resolution
67 lp/ mm we could easily detect microcalcications of
10 m in size (1012 lp/mm).
An additional attractive feature of this technique is the
transient nature in display of different spatial frequencies in
the Fourier spectrum of an object. Abnormalities detected in
mammography are classied as spiculated masses, stellate
lesions, circumscribed masses, and microcalcications. As
these abnormalities vary in size and intensity they corre-
spond to different spatial frequency bands in the Fourier
spectrum of the mammogram. From the inlet in Fig. 2 we
can infer that at a given time, all the spatial frequencies exist
with different diffraction efciencies. The maximum ef-
ciency occurs for a selected band of frequencies that are
optimized by matching the intensities of the object and ref-
erence beams. Other spatial frequencies also appear but are
lower intensity. The dominant frequency band persists for a
relatively longer time whereas the other frequencies decay
fast. Thus we can distinguish between different spatial fre-
quencies as they reveal at different times. We recorded a
movie of the whole hologram recording and readout process
using a fast CCD so that the readout process can revisit as
well as freeze the frames for the radiologist to concentrate on
a particular band of frequencies. To demonstrate this concept
we used a resolution chart (USAF negative target, Edmund
Optics). As shown in Fig. 4(a) at time t =0 we can observe all
the frequency groups [(A) low frequency group, (B) middle
frequency group, and (C) high frequency group] but at time
at t =5 only the frequency group B which matches the refer-
ence beam intensity remains clear while other frequency
groups vanish. The radiologist also has the choice of select-
ing the dominant frequency band by varying the reference
beam intensity during the whole process. This could be a
potential advantage to the radiologist in diagnostics of abnor-
malities in the mammogram that occur at different scales
such as clusters of microcalcications at small scales, edges
of smooth or star-shaped objects in the breast at middle
scales and architectural distortions at large scales.
14
In conclusion, interesting results on temporal and
intensity-dependent properties of Fourier holographic grat-
ings using bR lms are reported. The analysis using a theo-
retical model based on photoinduced isomerization agrees
with our experimental results. We exploit this technique to
process mammograms in real-time for detection of microcal-
cications buried in the background of soft tissue. The re-
sults are useful to radiologists for early diagnostics of breast
cancer. In addition, we found that the information in static
processed images can be further separated in time scales.
These temporal image features offer useful information to
the radiologist about the pathological changes in the image
which cannot be obtained from the static processed image.
This optical technique can be easily adopted for processing
digital mammograms by replacing the object in Fig. 1 with
SLM connected to the computer.
This work is supported by National Cancer Institute,
NIH, Grant No. 1R21CA89673-01A1. The high quality pho-
tographic images of the clinical mammograms were obtained
from the University of Massachusetts Medical School,
Worcester, MA. We thank Dr. Carl DOrsi and Professor An-
drew Karellas for providing them. We are grateful to Dr.
Joby Joseph, I. I. T. Delhi, India, for valuable discussions.
1
J. F. Heanue, M. C. Bashaw, and L. Hesselink, Science 265, 749 (1994).
2
J. W. Goodman, Opt. Photonics News 2, 11 (1991).
3
R. H. Berg, S. Hvilsted, and P. Ramanujam, Nature (London) 383, 505
(1996).
4
J. P. Huignard and J. P. Herriau, Appl. Opt. 17, 2671 (1978).
5
T. Y. Chang, J. H. Hong, and P. Yeh, Opt. Lett. 15, 743 (1990).
6
D. Dantsker, J. Nonlinear Opt. Phys. Mater. 5, 775 (1996).
7
P. Wu, D. V. G. L. N. Rao, B. R. Kimball, M. Nakashima, and B. S.
DeCristofano, Appl. Phys. Lett. 81, 3888 (2002).
8
J. Joseph, F. J. Arnanda, D. V. G. L. N. Rao, J. A. Akkara, and M.
Nakashima, Opt. Lett. 21, 1499 (1996).
9
R. R. Birge, Sci. Am. 3, 90 (1995).
10
R. Thoma, N. Hampp, C. Brauchle, and D. Oesterhelt, Opt. Lett. 16, 651
(1991).
11
Z. Yuan, Y. Bao-Li, W. Ying-Li, M. Neimule, L. Ming, C. Guo-Fu and N.
Hampp, Chin. Phys. Lett. 20, 671 (2003).
12
A. Lewis, Y. Albeck, Z. Lange, J. Benchowski, and G. Weizman, Science
275, 1462 (1997).
13
P. Wu, L. Wang, J. Xu, B. Zou, X. Gong, G. Zhang, G. Tang, and W. Chen,
Phys. Rev. B 57, 3874 (1998).
14
H. S. Wong, L. Guan, and H. Kong, Proceedings of the 1995 International
Conference on Image Processing, ICIP 95, 21.
FIG. 3. Experimental results of image processing using Fourier holography.
For the binary image E: (a) original, (b) edge-enhanced, and (c) edge-
softened. For the mammograms: (d) original image with of region of interest
(ROI) circled by radiologists, (e) blowup of the ROI; (f) processed image
showing calcications.
FIG. 4. Transient display of spatial frequency information of grating reso-
lution chart captured at times (a) t =0 and (b) t =5 s.
5838 Appl. Phys. Lett., Vol. 85, No. 24, 13 December 2004 Kothapalli et al.
Downloaded 29 Mar 2012 to 158.121.194.143. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions